“Data-Driven Thinking” is a column written by members of the media community and containing fresh ideas on the digital revolution in media.
Today’s column is written by Mark Hughes, CEO of C3 Metrics.
At the worst possible time, the digital marketing measurement business is again showing inconsistencies. The issue this time is the matter of “viewable impressions,” and I’m taking issue with the information released last week by comScore. Whether we like it or not, Internet users are not seeing all served ads, and we need to find the most accurate way of painting this picture, which admittedly is not pretty.
Here’s the quick background. One of the key findings delivered last week by a comScore study of 12 big brands is that 31% of the 1.7 billion ad impressions sampled recently in its study were never in view. The issue is that comScore is taking 12 big brands with huge budgets, and then sampling them on premium sites only. It would be nice if real companies could buy, plan and measure media that way. But it’s not reality. It’s like calling Palo Alto, CA and Greenwich, CT a true representation of America. Not accurate, and not projectable for the large majority of advertisers.
And Nielsen? Nielsen just told Adexchanger in response to comScore’s data that visibility of ads only makes up 10-15% of the importance in digital advertising. Come again? If ads didn’t show on TV, but advertisers were charged for those ads–don’t you think there would be some immediate fire and brimstone meetings?
Historically, all server requests recorded by ad servers have been deemed impressions. However, the IAB, ANA and 4As have recently called for a “viewable impression” standard, which outlines that ads both fully load and enter into the viewable space on a consumer’s screen for at least one second.
The C3 Metrics Labs report found that 68% of all display ads served are never ultimately seen by consumers according to the viewable impression standard. Of the ads that are not seen, 12% never fully load. Because click-through rates are calculated based on ads that are actually seen, CTRs that do not take viewable impression data into account may be off by as much as 179%.
These numbers from C3 Metrics align with RealVu, the only ad server accredited by the Media Rating Council for a “viewable impression,” which has found that “up to 90% of ads on major networks never appear …. the best sites in the world only show a view rates of 75%.” [from Troy Tribe, President & COO, RealVu] Bottom line: comScore’s 30% number is a continent away from RealVu and C3 Metrics.
I would like to see two actions result from this divergent discrepancy. First, click-through rates must be calculated on impressions viewed, not server requests. Second, attribution modeling, which is critical to display success, must also incorporate the viewable impression standard, because view-through credit is entirely based on an actual view. As a result, CPMs for viewable impressions will justifiably rise, and unviewed impressions will no longer count in display ad performance reporting.
The viewable impression standards have arrived. There’s good news and bad news whenever a new standard is adopted, but the IAB, ANA, and 4As have pushed this standard, and ultimately more dollars will pour into digital advertising with new accountability that CMOs and CFOs are looking for. Any agency failing to adopt the new standard might be optimizing with click-through rates off by 179%, and from attribution modeling standpoint, will have results off by 68%.
Follow C3 Metrics (@C3metrics) and AdExchanger.com (@adexchanger) on Twitter.
